Would it be possible, that you now (with optimized configuration) run some more speed tests?
Could you please post the picture how you connected pigtails to diplexers and cards as well as content of /etc/config/wirelless
I am not sure if it could make the difference but basicaally something tell me it could. Maybe @hagrid could give us more details how to connect everything together.
I was also unable to switch-branch to hbd I get lot of errors that seems unrecoverable during pkgupdate after swith from TOS 5.3 to HBD. My plan is to to import medkit of TOS6 into factory snapshot and then do fresh reconfiguration. Of course will prefer to do it on more stable branch but it is still uncertain when the transiton will take the place as problem of kernel on PPC architecture is most likely still not resolved. Maybe @Pepe could give us some schedule
Cabling looks good even though I don’t understand why you use that single additional antenna. Best is to use identical antenna-configuration for all connections of the same Wi-Fi-card (either eight antennas or four antennas combined by diplexers). Even though there might only be small differences some devices could behave funny because of that.
well in fact i dont have knowledge enough for that. Single pigtail antenna is original one, so i kept it there and tried to use it. I did not expect, that its better to use both pigtails on 4 antennas, then to use 5 antennas like i have it. But is not big deal, if it would be better to have really configuration of 2x wifi cards + 4x diplexers + 4x antennas, i can reconnect it and remove completely that original single pigtail antenna. I have separated pigtale cable to do that.
So you think i should remove that and really use 4 antennas with 4 diplexres completely and to not use 5th antenna?
There are small numbers on the diplexers, near the antenna connectors, hardly visible, but they are there (if I remember - 2,4 and 5,1).
2,4 is for 2,4GHz card antennas and 5,1 for 5GHz card antennas.
It is very important to connect card to antennas right via diplexers, because each “path” from connector to antenna is optimized for that frequency!
Yes, absolutely. It may improve connectivity.
Do you see what i see?
On right is original diplexer, on left is new diplexer for turris from eshop.
Yes exactly, this is what I needed to see.
What we have to see? Diplexers are little bit different, but the numbers at connectors are still there (2.4 , 5.8)
yeahh but they are opposite, then on original diplexers.
Ok, but there are still those numbers, so no problem.
yes correct, but its from eshop, where image looks like original diplexer, is labeled as " Turris Omnia diplexer", but anyway delivered is evidently different diplexer, where pigtails are vice versa.
I know it works, but looks strange.
Maybe it is just “updated” version and all new Omnias have the same diplexers. I haven´t checked mine new diplexers, which I bought for these new AX cards, but I think they will be the same as yours.
The only place where it is possible to buy these diplexers is Discomp and few CZ and SK e-shops which have them from the same manufacturer (my opinion). So, there is no “fake” diplexer, I think.
Yes, it works, i dont have a problem with that at all, is just confusing. Anyway, i did send an email to discomp, if it is like that and if there is a some reason to have it differently (i have no idea, why i would create opposite design of diplexer, at least from cost point of view it has no sense to change something that works properly). I can imagine, what you told, original diplexers delivered different vendor, then current diplexers are created by another company, even if sold under one label from same shop.
I think, they made som optimisation, and this new is cheaper, or it has to be changed due the lack of semiconductors etc.